kernel.git/drivers/net/wireless/ath, branch linux-6.15.y

wifi: ath12k: install pairwise key first

2025-08-20T16:36:35Z

commit 66e865f9dc78d00e6d1c8c6624cb0c9004e5aafb upstream. As station, WCN7850 firmware requires pairwise key to be installed before group key. Currently host does not care about this, so it is up to kernel or userspace to decide which one will be installed first. In case above requirement is not met, WCN7850 firmware's EAPOL station machine is messed up, and finally connection fails [1]. Reorder key install for station interface in that case: this is done by caching group key first; Later when pairwise key arrives, both can be installed in required order. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.3.1-00217-QCAHKSWPL_SILICONZ-1 Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218733 Link: https://lore.kernel.org/all/AS8P190MB12051DDBD84CD88E71C40AD7873F2@AS8P190MB1205.EURP190.PROD.OUTLOOK.COM # [1] Signed-off-by: Baochen Qiang Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250523-ath12k-unicast-key-first-v1-2-f53c3880e6d8@quicinc.com Signed-off-by: Jeff Johnson Signed-off-by: Baochen Qiang Signed-off-by: Greg Kroah-Hartman

wifi: ath12k: Decrement TID on RX peer frag setup error handling

2025-08-20T16:36:11Z

[ Upstream commit 7c0884fcd2ddde0544d2e77f297ae461e1f53f58 ] Currently, TID is not decremented before peer cleanup, during error handling path of ath12k_dp_rx_peer_frag_setup(). This could lead to out-of-bounds access in peer->rx_tid[]. Hence, add a decrement operation for TID, before peer cleanup to ensures proper cleanup and prevents out-of-bounds access issues when the RX peer frag setup fails. Found during code review. Compile tested only. Signed-off-by: Karthikeyan Kathirvel Signed-off-by: Sarika Sharma Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250526034713.712592-1-quic_sarishar@quicinc.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath12k: Enable REO queue lookup table feature on QCN9274 hw2.0

2025-08-20T16:36:11Z

[ Upstream commit b79742b84e16e41c4a09f3126436f39f36e75c06 ] The commit 89ac53e96217 ("wifi: ath12k: Enable REO queue lookup table feature on QCN9274") originally intended to enable the reoq_lut_support hardware parameter flag for both QCN9274 hw1.0 and hw2.0. However, it enabled it only for QCN9274 hw1.0. Hence, enable REO queue lookup table feature on QCN9274 hw2.0. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ-1 Signed-off-by: Raj Kumar Bhagat Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250609-qcn9274-reoq-v1-1-a92c91abc9b9@quicinc.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath12k: Fix station association with MBSSID Non-TX BSS

2025-08-20T16:36:10Z

[ Upstream commit 70eeacc1a92a444f4b5777ab19e1c378a5edc8dd ] ath12k station is unable to associate with non-transmitting BSSes in a Multiple BSS set because the user-space does not receive information about the non-transmitting BSSes from mac80211's scan results. The ath12k driver does not advertise its MBSSID capability to mac80211, resulting in wiphy->support_mbssid not being set. Consequently, the information about non-transmitting BSS is not parsed from received Beacon/Probe response frames and is therefore not included in the scan results. Fix this by advertising the MBSSID capability of ath12k driver to mac80211. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ-1 Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00284-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1 Signed-off-by: Hari Chandrakanthan Signed-off-by: Rameshkumar Sundaram Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250530035615.3178480-2-rameshkumar.sundaram@oss.qualcomm.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath12k: Add memset and update default rate value in wmi tx completion

2025-08-20T16:36:10Z

[ Upstream commit 9903c0986f782dfc511d7638b6f15fb6e8600cd3 ] When both AP/STA and monitor interfaces are enabled, ieee80211_tx_status() is invoked from two paths: the TX completion handler for data frames and the WMI TX completion handler for management frames. In the data path, the skb->cb is properly zeroed using memset, but in the WMI path, this step is missing. As a result, mac80211 encountered uninitialized (junk) values in skb->cb when generating the radiotap header for monitor mode, leading to invalid radiotap lengths. Hence, explicitly zero the status field in the skb->cb using memset in WMI TX completion path to ensure consistent and correct behavior during WMI tx completion path. Additionally, set info->status.rates[0].idx = -1 to indicate that no valid rate information is available, avoiding misinterpretation of garbage values. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ-1 Signed-off-by: Sarika Sharma Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250603063512.1887652-1-quic_sarishar@quicinc.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath10k: shutdown driver when hardware is unreliable

2025-08-20T16:36:10Z

[ Upstream commit c256a94d1b1b15109740306f7f2a7c2173e12072 ] In rare cases, ath10k may lose connection with the PCIe bus due to some unknown reasons, which could further lead to system crashes during resuming due to watchdog timeout: ath10k_pci 0000:01:00.0: wmi command 20486 timeout, restarting hardware ath10k_pci 0000:01:00.0: already restarting ath10k_pci 0000:01:00.0: failed to stop WMI vdev 0: -11 ath10k_pci 0000:01:00.0: failed to stop vdev 0: -11 ieee80211 phy0: PM: **** DPM device timeout **** Call Trace: panic+0x125/0x315 dpm_watchdog_set+0x54/0x54 dpm_watchdog_handler+0x57/0x57 call_timer_fn+0x31/0x13c At this point, all WMI commands will timeout and attempt to restart device. So set a threshold for consecutive restart failures. If the threshold is exceeded, consider the hardware is unreliable and all ath10k operations should be skipped to avoid system crash. fail_cont_count and pending_recovery are atomic variables, and do not involve complex conditional logic. Therefore, even if recovery check and reconfig complete are executed concurrently, the recovery mechanism will not be broken. Tested-on: QCA6174 hw3.2 PCI WLAN.RM.4.4.1-00288-QCARMSWPZ-1 Signed-off-by: Kang Yang Reviewed-by: Loic Poulain Link: https://patch.msgid.link/20250623022731.509-1-kang.yang@oss.qualcomm.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath12k: Correct tid cleanup when tid setup fails

2025-08-20T16:36:02Z

[ Upstream commit 4a2bf707270f897ab8077baee8ed5842a5321686 ] Currently, if any error occurs during ath12k_dp_rx_peer_tid_setup(), the tid value is already incremented, even though the corresponding TID is not actually allocated. Proceed to ath12k_dp_rx_peer_tid_delete() starting from unallocated tid, which might leads to freeing unallocated TID and cause potential crash or out-of-bounds access. Hence, fix by correctly decrementing tid before cleanup to match only the successfully allocated TIDs. Also, remove tid-- from failure case of ath12k_dp_rx_peer_frag_setup(), as decrementing the tid before cleanup in loop will take care of this. Compile tested only. Signed-off-by: Sarika Sharma Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250721061749.886732-1-quic_sarishar@quicinc.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath12k: fix endianness handling while accessing wmi service bit

2025-08-15T10:16:33Z

[ Upstream commit 8f1a078842d4af4877fb686f3907788024d0d1b7 ] Currently there is no endian conversion in ath12k_wmi_tlv_services_parser() so the service bit parsing will be incorrect on a big endian platform and to fix this by using appropriate endian conversion. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00217-QCAHKSWPL_SILICONZ-1 Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3 Fixes: 342527f35338 ("wifi: ath12k: Add support to parse new WMI event for 6 GHz regulatory") Signed-off-by: Tamizh Chelvam Raja Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250717173539.2523396-2-tamizh.raja@oss.qualcomm.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath11k: fix sleeping-in-atomic in ath11k_mac_op_set_bitrate_mask()

2025-08-15T10:16:31Z

[ Upstream commit 65c12b104cb942d588a1a093acc4537fb3d3b129 ] ath11k_mac_disable_peer_fixed_rate() is passed as the iterator to ieee80211_iterate_stations_atomic(). Note in this case the iterator is required to be atomic, however ath11k_mac_disable_peer_fixed_rate() does not follow it as it might sleep. Consequently below warning is seen: BUG: sleeping function called from invalid context at wmi.c:304 Call Trace: dump_stack_lvl __might_resched.cold ath11k_wmi_cmd_send ath11k_wmi_set_peer_param ath11k_mac_disable_peer_fixed_rate ieee80211_iterate_stations_atomic ath11k_mac_op_set_bitrate_mask.cold Change to ieee80211_iterate_stations_mtx() to fix this issue. Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30 Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices") Signed-off-by: Baochen Qiang Link: https://patch.msgid.link/20250603-ath11k-use-non-atomic-iterator-v1-1-d75762068d56@quicinc.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin

wifi: ath12k: Use HTT_TCL_METADATA_VER_V1 in FTM mode

2025-08-15T10:16:31Z

[ Upstream commit 66b3ebc77d23d6574a965bdbfe41de8aeb7f384e ] Currently host sends HTT_TCL_METADATA_VER_V2 to the firmware regardless of the operating mode (Mission or FTM). Firmware expects additional software information (like peer ID, vdev ID, and link ID) in Tx packets when HTT_TCL_METADATA_VER_V2 is set. However, in FTM (Factory Test Mode) mode, no vdev is created on the host side (this is expected). As a result, the firmware fails to find the expected vdev during packet processing and ends up dropping packets. To fix this, send HTT_TCL_METADATA_VER_V1 in FTM mode because FTM mode doesn't support HTT_TCL_METADATA_VER_V2. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.5-01651-QCAHKSWPL_SILICONZ-1 Fixes: 5d964966bd3f ("wifi: ath12k: Update HTT_TCL_METADATA version and bit mask definitions") Signed-off-by: Aaradhana Sahu Reviewed-by: Vasanthakumar Thiagarajan Link: https://patch.msgid.link/20250711035420.1509029-1-aaradhana.sahu@oss.qualcomm.com Signed-off-by: Jeff Johnson Signed-off-by: Sasha Levin